**** BEGIN LOGGING AT Sat Jul 23 23:59:56 2005 Jul 24 00:41:44 berlios.de seems to be back up. Jul 24 00:55:48 but we have migrated bitbake away from it anyway. We now keep our own copy of the bitbake sources in nslu2-linux.org so that berlios.de going down doesn't stop us building. Jul 24 00:59:01 03rwhitby 07org.nslu2-linux.dev * rc7aaea35... 10/Makefile: Migrated bitbake from the unstable and three-day-downtime berlios.de SVN to our well-maintained nslu2-linux.org monotone repository Jul 24 01:01:26 rwhitby-away: Good move Jul 24 01:30:02 Now we just need to implement jbowler's BBFILES scheme and we can remove the build dependency on berlios.de altogether, seeing as they can't keep their anonymous SVN service running reliably ... Jul 24 01:34:11 Hmm, I seem to have broken slugbot, hes not logging to loglibrary anymore. Jul 24 01:36:37 foo Jul 24 01:38:04 ok thats better. Jul 24 02:44:04 jp30-work: I also get the following optware build error on nudi: Jul 24 02:44:09 (cd /home/slug/optware/builds; cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/byrequest co -D "2005-04-23" byRequest;) Jul 24 02:44:10 cvs [checkout aborted]: unrecognized auth response from cvs.sourceforge.net: M PserverBackend::PserverBackend() Connect (Connection refused) Jul 24 03:58:02 Got errors in making glibc for Unslung: Jul 24 03:58:04 /home/slug/nslu2/unslung/tmp/work/glibc-2.2.5-r5/build-armeb-linux/csu/crti.S:96 Jul 24 03:58:06 : Error: can't resolve `_GLOBAL_OFFSET_TABLE_' {*UND* section} - `.L6' {.text se Jul 24 03:58:07 ction} Jul 24 03:58:09 [more errors follow] Jul 24 03:58:10 I got this on the latest monotone, svn, cvs after wiping /home/slug/nslu2/unslung/tmp/work/glibc-2.2.5-r5 and corresponding stamps after getting the error for the first time. Jul 24 04:08:25 bwalle2: Your Streamripper patch has now made it to Openslug. Jul 24 04:08:52 Cool Jul 24 04:09:37 I've run it for a while on unslung now - definitely much better at getting complete track rips. Jul 24 04:22:01 dyoung-away, testing Jul 24 04:22:15 foo Jul 24 04:22:18 nefarious Jul 24 04:22:21 tee hee Jul 24 04:22:36 244 Jul 24 04:28:20 ☂ Jul 24 05:55:30 Is there any technical reason why gcj (the Java compiler) is not part of the cross-toolchain? Jul 24 08:23:52 ingeba-toandfrom, ping Jul 24 08:41:46 rats Jul 24 10:06:47 03bzhou * 10unslung/ (4 files in 2 dirs): cherokee built Jul 24 10:42:41 03bzhou * 10unslung/make/cherokee.mk: use pthread Jul 24 13:27:14 03nail 07org.openembedded.nslu2-linux * rf3173306... 10/packages/ncftp/ncftp_3.1.9.bb: Upstream upgrade Jul 24 13:27:17 03nail 07org.openembedded.nslu2-linux * r5b948a2a... 10/packages/man/ (man/man.conf man_1.5p.bb): Upstream upgrade Jul 24 13:27:19 03nail 07org.openembedded.nslu2-linux * r9203ea8d... 10/packages/streamripper/streamripper_1.61.10.bb: Added streamripper Jul 24 13:30:32 03nail 07org.openembedded.nslu2-linux * rf3826b85... 10/packages/groff/ (groff/groff.patch groff_1.19.1.bb): Added. Needed for man Jul 24 13:30:35 03nail 07org.openembedded.nslu2-linux * rf9ab42d4... 10/packages/procps/ (5 files in 2 dirs): updated version, patched to support OE cross install-sh Jul 24 13:34:49 03repvik * r95 10/trunk/openslug/nslu2-linux/packages/ (groff man ncftp procps streamripper): Added ncftp, streamripper, man, groff (needed for man) and procps. Jul 24 13:39:40 03repvik * r96 10/trunk/openslug/nslu2-linux/packages/man-pages: Added man-pages Jul 24 13:45:04 03nail 07org.openembedded.nslu2-linux * r2bdbe69f... 10/ (27 files in 19 dirs): Jul 24 13:45:04 propagate from branch 'org.openembedded.dev' (head 566032cf08a6d55f47a8f64bc3e5910a0ba4c0cd) Jul 24 13:45:04 to branch 'org.openembedded.nslu2-linux' (head f9ab42d4a7da6feeab5ad848a3bfad4b022a5a53) Jul 24 13:49:33 03repvik * r97 10/trunk/openslug/nslu2-linux/packages/screen: Added screen Jul 24 13:52:50 03repvik * r98 10/trunk/openslug/nslu2-linux/packages/psmisc: Added psmisc for testing Jul 24 13:54:39 03nail 07org.openembedded.nslu2-linux * rcacdd0f9... 10/packages/screen/ (screen-4.0.2/configure.patch screen_4.0.2.bb): Added screen Jul 24 13:57:31 ~emulate rwhitby Jul 24 13:58:35 hi Jul 24 13:59:26 ohh, things moved on :) (... wheres the end-user channel?) Jul 24 13:59:40 #nslu2-general Jul 24 14:03:44 * mumptai should better use his eyes Jul 24 14:10:11 we used an unslung image to control a autonomous robot early this year, i havn't visited this channel or the wiki since that time. i see a lot of new thing and improvements Jul 24 14:10:29 congratulations, fine work ;) Jul 24 14:10:54 Yup. I'm not working on unslung, but I see there's a lot of work being done Jul 24 14:11:51 that end-user seperation seems to have become important Jul 24 14:12:30 well, getting OE up an running did that quiet well before Jul 24 14:12:45 There's a fair bit of noise if it isn't separated, and endusers don't need to know the dirty details :P Jul 24 14:14:14 signal-to-noise ratio is important ... Jul 24 14:14:27 yeah Jul 24 14:15:16 okay, i don't want to make it worse. just wandered around the irc and thought i could say hello and say thanks Jul 24 14:15:35 hehe, no problems and thanks for the feedback Jul 24 14:18:59 if i come back in september/october i'm building the next robot and am trying to get up to date Jul 24 14:19:11 so long and thanks for the code Jul 24 14:28:13 NAiL, ping Jul 24 14:28:46 ingeba-away, ping Jul 24 14:29:02 ka6sox: ping Jul 24 14:29:38 ka6sox: pong Jul 24 14:30:03 hi ingeba-away I have a few questions if you have some time. Jul 24 14:30:18 ka6sox: Shoot - pub or priv as you prefer Jul 24 14:35:58 ncftp, man, groff, procps, streamripper, screen and psmisc need to be added to slugbug ;) Jul 24 14:38:54 NAiL: Thanks Jul 24 14:41:42 oh, right, I think I need to be added to slugbug, forgot about that. :) Jul 24 14:44:12 NAiL: man-pages ? Kind of useless with man without them. Jul 24 14:44:38 yes :) Jul 24 14:45:20 hi all, I updated my old monotone setup on my home machine and I now get errors wile parsing openslug/oe-symlinks/packages/*/*.bb Jul 24 14:46:20 Anyone owning "screen"? Jul 24 14:46:35 running bb openslug-image returns: No such file or directory: '/blah/blah/openslug/oe-symlinks/packages/*/*.bb' while parsing /blah/blah/openslug/oe-symlinks/packages/*/*.bb ~ Jul 24 14:46:40 monotone id please.... Jul 24 14:46:43 anyone seen something like this Jul 24 14:47:19 the files are there, as text files with link commands inside Jul 24 14:47:34 ingeba-away: no Jul 24 14:47:42 ok Jul 24 14:48:31 hanjo_: I think you are using an too old svn.. Jul 24 14:48:38 the files should be links Jul 24 14:49:02 NAiL: Added Jul 24 14:49:32 ingeba-away: thanks a bunch Jul 24 14:49:35 DaKa2: too old svn, as in the tool ? Jul 24 14:49:47 hanjo_: yes, the svn client Jul 24 14:50:19 thx, I'll try to upgrade Jul 24 14:57:45 DaKa2: that was it, thx for the tip Jul 24 14:58:07 :-) Jul 24 14:59:34 <[g2]> hey hanjo what got fixed ? Jul 24 14:59:53 old svn client Jul 24 15:00:04 <[g2]> ah... Jul 24 15:00:06 no links in oe-symlinks, just files with link info :-) Jul 24 15:00:25 <[g2]> hanjo, so what the status with nfsv3+tcp client/server ? Jul 24 15:00:57 well, I have tested nfs3 with tcp, but it takes almost the same time as UDP Jul 24 15:01:25 there seems to be some major problem with fc3 nfs server and ext3, so my problems are on the server side Jul 24 15:01:55 <[g2]> have you run nfsv3+tcp server from the slug ? Jul 24 15:01:59 It would have been cool to test nfs4. But I could not find a way to make busybox mount speek v4 Jul 24 15:02:07 g2: no Jul 24 15:02:21 the server is a FC3 box Jul 24 15:02:28 <[g2]> ok... Jul 24 15:02:32 the slugs are now at beta-2.3 Jul 24 15:03:06 <[g2]> is the fix for the nfs turnup in the repo yet ? Jul 24 15:03:08 So the server returns a bunch of "file does not exist" errors Jul 24 15:03:34 <[g2]> or is that a non-problem ? Jul 24 15:03:51 the /dev/root in rw ? Jul 24 15:04:13 <[g2]> I think that was one Jul 24 15:04:20 that is fixed. shutdown -r works fine now Jul 24 15:04:30 <[g2]> cool Jul 24 15:05:57 yea, If only the unionfs guys get their act together and finaly do the persistent inodes trick, it will be a very elegant setup Jul 24 15:08:50 <[g2]> what's that unionfs mount rw of the nfs root ? Jul 24 15:09:16 <[g2]> s/of/over Jul 24 15:12:14 the problem is that the slugs are locking if the nfs server restarts because the exported file system from union fs uses virtual inodes Jul 24 15:12:49 "monotone: misuse: branch 'org.nslu2-linux.bitbake' is empty" :-( Jul 24 15:12:49 the same thing happens if you modify a folder under the union Jul 24 15:13:40 With persistent inodes, unionfs will always maintain the mapping between the exported uniion and the real inodes on the disk, so these issues will be resolved Jul 24 15:19:12 ingeba-away: failed to build ncftp and groff Jul 24 15:19:31 NAiL: Really? Jul 24 15:19:37 yup Jul 24 15:20:09 /usr/bin/install -c -m 644 img/pic* /home/repvik/unstable/glibc/openslug/tmp/work/groff-1.19.1-r0/image/usr/share/doc/groff/1.19.1/html/img Jul 24 15:20:13 /usr/bin/install: cannot stat `img/pic*': No such file or directory Jul 24 15:20:17 That's from groff Jul 24 15:20:31 -> /usr/bin/install ???? Jul 24 15:20:39 apparently Jul 24 15:20:55 Where on earth did that come from? rebuilding groff here now Jul 24 15:21:03 Flawless Jul 24 15:21:38 hmm.. ncftp built now Jul 24 15:21:40 odd Jul 24 15:21:52 which -a install; install --version Jul 24 15:22:17 /usr/bin/install Jul 24 15:22:17 install (coreutils) 5.2.1 Jul 24 15:22:20 (thought it would be useing the staged install-sh in i686 anyway Jul 24 15:22:30 Same here (Gentoo) Jul 24 15:23:07 debian etch Jul 24 15:23:38 What about ncftp? Jul 24 15:23:51 it built now... strange Jul 24 15:24:08 Del stamps and work dirs. Jul 24 15:24:35 Both builds here Jul 24 15:24:42 (from scratched work dirs) Jul 24 15:25:05 (Can't build unslung glibc though - broken since of today) Jul 24 15:25:46 NAiL: Gonna go to sleep. Can you mail me the runs and logs from the latest failed build of each failure? Jul 24 15:26:12 Oh - lets' switch to openslug - I have a question. Jul 24 15:36:24 do we have binfmt_misc on unslung 5? it is not in 3.18 Jul 24 15:38:41 DaKa2: make setup will update your branch include list Jul 24 15:39:06 ah Jul 24 15:39:46 rwhitby: FYI - About to go to sleep, but after a full remake attempt of Unslung from a fresh pull still results in failure to create glibc. Just keep in mind in case there are more reports Jul 24 15:39:59 thanks, finally a working make update Jul 24 15:40:41 ingeba-zzz: there were reports of glibc problems in #oe yesterday Jul 24 15:45:13 03nail 07org.openembedded.nslu2-linux * r784c0507... 10/packages/meta/openslug-packages.bb: Added man, man-pages, ncftp, psmisc, screen and streamripper Jul 24 15:46:02 My work here is done ;-) Jul 24 15:49:16 ~emulate rwhitby Jul 24 15:49:47 "Another Satisfied Customer" Jul 24 15:51:51 DaKa2: no update-alternatives problems reported by #oe people yet - a great testament to the quality of your work! Jul 24 15:52:00 :) Jul 24 15:52:08 only one reported from here :-) Jul 24 15:52:10 and fixed Jul 24 15:54:58 03daka 07org.openembedded.nslu2-linux * r5bbc6ec4... 10/packages/meta/openslug-native.bb: Add gcc-symlinks, fix the files libpthread.so and libc.so from libc6-dev (that really should be fixed in there) Jul 24 15:55:27 whoops.. thats wrong.. actually added ccp-symlinks, I hope.. Jul 24 15:55:30 cpp* Jul 24 16:07:08 DaKa2: can you take up the discussion of the libc6 fixes with mickeyl & co ? Jul 24 16:07:53 Yep, I could try :-) Im looking into where and why its put there right now Jul 24 16:08:11 thx Jul 24 17:13:03 03jp30 * 10unslung/ (3 files in 2 dirs): finishing touches to qemu, use crosstool to build a runtime environment for qemu-i386 Jul 24 17:16:14 03jp30 * 10unslung/Makefile: make elements of unslung's cross-configuration available as makefile variables Jul 24 18:13:05 03bzhou * 10unslung/make/bluez-hcidump.mk: upstream upgrade from 1.22 to 1.24 Jul 24 18:14:04 03bzhou * 10unslung/make/ (bluez-libs.mk bluez-utils.mk): upstream upgrade from 2.17 to 2.18 Jul 24 18:31:38 03jp30 * 10unslung/make/ (sqsh.mk freetds.mk): sqsh: depends on ncurses; freetds: those libtool archives contain host paths Jul 24 18:37:30 jp30, thx for fixing freetds and sqsh problems Jul 24 18:40:12 jp30: ping Jul 24 18:40:50 jp30: (cd /home/slug/optware/builds; cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/byrequest co -D "2005-04-23" byRequest;) Jul 24 18:40:58 cvs [checkout aborted]: unrecognized auth response from cvs.sourceforge.net: cvs pserver: cannot read /cvsroot/byrequest/CVSROOT/config: Input/output error Jul 24 18:41:03 any idea? Jul 24 18:44:02 eno-away, no probs. turns out bluez needs some fixing too - these are just errors that unslung-check-packages.pl turned up. if you link with .la files in staging, typically you get -rpath=$(STAGING_LIB_DIR) which is bad Jul 24 18:44:11 rwhitby-web, pong Jul 24 18:45:18 jp30: are you currently able to build optware from scratch on nudi? Jul 24 18:45:40 hmm... shall try a rebuild of byRequest and see what happens? Jul 24 18:45:59 i will just commit my current work on bluez first Jul 24 18:46:08 ok, no hurry :-) Jul 24 18:47:10 I'm just still trying to get a hat-trick of openslug, unslung and optware all building on nudi at a single point in time. It hasn't been possible for some time for me .... Jul 24 18:48:09 03jp30 * 10unslung/make/bluez-libs.mk: don't stage libtool archives Jul 24 18:48:52 and now glibc is failing with a bitbake error in unslung firmware, so that's only 1 out of 3 at the moment ... Jul 24 18:55:18 rwhitby-web: looks like it is sourceforge's problem. still, it was probably bad practice not to cache cvs-downloaded sources Jul 24 18:56:24 yeah, its done with a date, so we could cache it like we do in puppy Jul 24 18:58:28 yeah, i will update template-cvs.mk to make sure it encourages developers to do that Jul 24 18:59:54 rwhitby-web, do you know the answer to the unslung firmware question i posted earlier? does unslung 5 have binfmt_misc compiled into the kernel / available as a module? it would make qemu work a lot smoother. Jul 24 19:00:45 I don't know what binfmt_misc is, so the answer is likely to be no. Jul 24 19:02:19 ok, i'll add it to the appropriate wish list. it is a kernel feature that allows the interpreters to be registers for binaries with some simple magic number matching. you can configure it so it knows to use qemu to interpret a i386 elf binary Jul 24 19:03:21 on desktops, it's often used to set wine as the interpreter for microsoft .exe binaries Jul 24 19:04:10 03daka 07org.openembedded.nslu2-linux * r2eeab934... 10/packages/meta/openslug-native.bb: Add binutils-symlinks, remove the libc.so and libpthread.so fix since things work without it Jul 24 19:12:36 03jp30 * 10unslung/Makefile: promote freetds sqsh py-mssql qemu qemu-libc-i386 Jul 24 19:20:42 jp30: we need to get you making kernel mods so that you can add the things you need for the packages. From now on I expect it will be package needs that drive the Unslung firmware. Jul 24 19:27:19 rwhitby, i'm ready to try messing with the firmware. i could do with some hand-holding learning to use monotone though Jul 24 19:29:33 my attempts to get started using the master makefile have tended to end with some error involving berlios being on the blink. let me try again now Jul 24 19:30:34 yeah, it should work now, and I have a selfish interest in helping you get up and running with firmware mods :-) Jul 24 19:31:39 ok, it's sitting on "verifying new revisions" at the moment - looks like this will take a few minutes Jul 24 19:32:19 the process of doing a pull with monotone seems very processor-heavy - is that normal? Jul 24 19:33:01 hey, did i mention i accepted that job offer i was talking about a while ago. i move to NZ in January Jul 24 19:33:06 yes. and it may take more than a few minutes Jul 24 19:33:22 hey - congrats! now we'll be in closer timezones! Jul 24 19:34:12 27 revs / 90; yeah, i'll have to look at new power supplies for my slugs too Jul 24 19:36:29 congrats jp30! Jul 24 19:36:42 jp30: I just bought a UPS for slug-central (http://nslu2-linux.thestuffguy.com/gallery/slug-central/IMGP4012) - a 600VA model runs that lot for 40 minutes. Jul 24 19:40:53 it is one of the "Smart" UPSes, right? Jul 24 19:41:59 www.upsonic.com.au - PrOffice 600VA Jul 24 19:42:32 thanks ka6sox Jul 24 19:42:39 purchase decision was mainly based on form-factor - the UPS now hides the power points so that little fingers can't get to them. Jul 24 19:42:55 aha... Jul 24 19:44:16 well just from experience (and I sell UPSes) the "Smart" UPSes have extra internals that allow them to better handle brown outs and slight over-voltages - which tend to kill the lifetime of standard "Back" UPSes Jul 24 20:00:13 on that product - PrOffice 600 - (from what little I could read before Acrobat decided to die) - just make sure that you test the UPS at least once a month by unplugging it Jul 24 20:01:07 that way, you might be able to anticipate the potential battery and/or battery charger death (which is the first thing to go in standard "Back" UPSes) Jul 24 20:02:34 it has a self-test feature which turns of the AC and runs until battery low. Jul 24 20:02:50 I need to get that running under control from the slug. Jul 24 20:02:58 hehe Jul 24 20:03:17 I smell a package in process Jul 24 20:14:04 03daka 07org.openembedded.nslu2-linux * r27ad08fd... 10/packages/busybox/busybox_1.00.bb: Jul 24 20:14:04 Check for existance of 'test' before assuming it isn't there and alias test to 'busybox test'. Jul 24 20:14:04 When building a rootfs the previos implementation cased a few serious (non-fatal) errors showing up with tinderbox. Jul 24 20:14:39 rwhitby-web, i'm having trouble building unslung firmware - looks as though oe's libtool-native package is broken at the mo? Jul 24 20:44:09 jp30: my unslung build is broken too Jul 24 20:45:46 my libtool-native builds fine, but glibc breaks with a bitbake error Jul 24 21:49:20 03bzhou * 10unslung/make/getmail.mk: upstream upgrade from 4.3.6 to 4.3.11 Jul 24 22:04:11 03bzhou * 10unslung/make/mailman.mk: upstream upgrade from 2.1.5 to 2.1.6 Jul 24 22:23:44 03bzhou * 10unslung/ (3 files in 2 dirs): upstream upgrade from 1.7.8 to 1.7.9 Jul 24 22:36:51 03bzhou * 10unslung/make/py-roundup.mk: upstream upgrade from 0.8.3 to 0.8.4 **** ENDING LOGGING AT Sun Jul 24 23:59:56 2005